Skip to content

By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

Lenny's PodcastLenny's Podcast

Edwin Chen: Why optimizing for benchmarks creates AI sloth

How Surge bootstrapped past $1B revenue with fewer than 100 people; Chen argues benchmark gaming pushes AI toward dopamine, emojis, and slop, not truth.

Lenny RachitskyhostEdwin Chenguest
Dec 7, 20251h 10mWatch on YouTube ↗

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

Episode Details

EPISODE INFO

Released
December 7, 2025
Duration
1h 10m
Channel
Lenny's Podcast
Watch on YouTube
▶ Open ↗

EPISODE DESCRIPTION

Edwin Chen is the founder and CEO of Surge AI, the company that teaches AI what’s good vs. what’s bad, powering frontier labs with elite data, environments, and evaluations. Surge surpassed $1 billion in revenue with under 100 employees last year, completely bootstrapped—the fastest company in history to reach this milestone. Before founding Surge, Edwin was a research scientist at Google, Facebook, and Twitter and studied mathematics, computer science, and linguistics at MIT. *We discuss:*

  1. How Surge reached over $1 billion in revenue with fewer than 100 people by obsessing over quality
  2. The story behind how Claude Code got so good at coding and writing
  3. The problems with AI benchmarks and why they’re pushing AI in the wrong direction
  4. How RL environments are the next frontier in AI training
  5. Why Edwin believes we’re still a decade away from AGI
  6. Why taste and human judgment shape which AI models become industry leaders
  7. His contrarian approach to company building that rejects Silicon Valley’s “pivot and blitzscale” playbook
  8. How AI models will become increasingly differentiated based on the values of the companies building them

*Brought to you by:* Vanta—Automate compliance. Simplify security: https://vanta.com/lenny WorkOS—Modern identity platform for B2B SaaS, free up to 1 million MAUs: https://workos.com/lenny Coda—The all-in-one collaborative workspace: https://coda.io/lenny *Transcript:* https://www.lennysnewsletter.com/p/surge-ai-edwin-chen *My biggest takeaways (for paid newsletter subscribers):* https://www.lennysnewsletter.com/i/180055059/my-biggest-takeaways-from-this-conversation *Where to find Edwin Chen:*

*Where to find Lenny:*

*In this episode, we cover:* (00:00) Introduction to Edwin Chen (04:48) AI’s role in business efficiency (07:08) Building a contrarian company (08:55) An explanation of what Surge AI does (09:36) The importance of high-quality data (13:31) How Claude Code has stayed ahead (17:37) Edwin’s skepticism toward benchmarks (21:54) AGI timelines and industry trends (28:33) The Silicon Valley machine (33:07) Reinforcement learning and future AI training (39:37) Understanding model trajectories (41:11) How models have advanced and will continue to advance (42:55) Adapting to industry needs (44:39) Surge’s research approach (48:07) Predictions for the next few years in AI (50:43) What’s underhyped and overhyped in AI (52:55) The story of founding Surge AI (01:02:18) Lightning round and final thoughts *Referenced:*

*Recommended books:*

_Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email podcast@lennyrachitsky.com._ Lenny may be an investor in the companies discussed.

SPEAKERS

  • Lenny Rachitsky

    host
  • Edwin Chen

    guest
  • Narrator

    other

EPISODE SUMMARY

In this episode of Lenny's Podcast, featuring Lenny Rachitsky and Edwin Chen, Edwin Chen: Why optimizing for benchmarks creates AI sloth explores Bootstrapped AI Data Giant Surge Reimagines Responsible Path To AGI Founder Edwin Chen explains how Surge AI became a $1B-revenue, sub‑100‑person, fully bootstrapped company by obsessing over ultra‑high‑quality training data for frontier models like ChatGPT, Claude, and Gemini.

RELATED EPISODES

Why securing AI is harder than anyone expected and guardrails are failing | HackAPrompt CEO

Why securing AI is harder than anyone expected and guardrails are failing | HackAPrompt CEO

The new AI growth playbook for 2026 | How Lovable hit $200M ARR in one year

The new AI growth playbook for 2026 | How Lovable hit $200M ARR in one year

Inside OpenAI: 2026 is the year of agents, AI’s biggest bottleneck, and why compute isn’t the issue

Inside OpenAI: 2026 is the year of agents, AI’s biggest bottleneck, and why compute isn’t the issue

Why AI is disrupting traditional product management | Tomer Cohen (LinkedIn CPO)

Why AI is disrupting traditional product management | Tomer Cohen (LinkedIn CPO)

What world-class GTM looks like in 2026 | Jeanne DeWitt Grosser (Vercel, Stripe, Google)

What world-class GTM looks like in 2026 | Jeanne DeWitt Grosser (Vercel, Stripe, Google)

A guide to difficult conversations, building high-trust teams, and designing a life you love

A guide to difficult conversations, building high-trust teams, and designing a life you love

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.